gensim: Utilities to interface to gensim.
In PolMine/polmineR.topics: tools for topicmodelling

Description Usage Arguments Details Author(s) Examples

Utilities to interface to gensim.

gensim_ldamodel_as_LDA_Gibbs(model, dtm)

gensim_ldamodel_load(modeldir, modelname)

dtm_as_bow(dtm)

`model`	An LDA model trained by gensim.
`dtm`	A Document-Term-Matrix (will be turned into BOW data structure).
`modeldir`	Directory where a gensim LDA topic model has been saved.
`modelname`	Name of a gensim LDA topic model. The data for a model consists of a set of files starting with the modelname each.

The gensim_ldamodel_as_LDA_Gibbs()-function turns a gensim/Python model (that may have been loaded using gensim_ldamodel_load) into the class LDA_Gibbs well-known from the topicmodels package for further processing within R.

Use gensim_ldamodel_load to load an ldamodel computed by gensim. The return value is a LdaModel Python object that can serve as input to functions or that can be processed using the reticulate package.

The input to gensim's LDA modelling methods is a representation of corpora in a data format denoted as "BOW". This utility function dtm_as_bow turns a sparse matrix (class simple_triplet_matrix) into the bow input format required by gensim.

Andreas Blaette

## Not run: 
if (requireNamespace("reticulate")){
  gensim <- reticulate::import("gensim")

  modeldir <- system.file(package = "topicanalysis", "extdata", "gensim")
  modelname <- "germaparlmini"
  
  dtm <- readRDS(
    file = system.file(
      package = "topicanalysis", "extdata", "gensim", "germaparlmini_dtm.rds"
    )
  )
  
  lda <- gensim_ldamodel_load(
    modeldir = system.file(package = "topicanalysis", "extdata", "gensim"),
    modelname = "germaparlmini"
  )

  y <- gensim_ldamodel_as_LDA_Gibbs(model = lda, dtm = dtm)
  topics_terms <- topicmodels::get_terms(y, 10)
  docs_topics <- topicmodels::get_topics(y, 5)
}

## End(Not run)